VP, AI Automation Engineer

Customers Bank•Malvern, PA

About The Position

The AI Automation Engineer is a builder and owner of the Bank’s intelligent automation layer — designing agentic pipelines, multi-step AI workflows, and the prompt and evaluation frameworks that keep model outputs production-grade. This is a high-impact, execution-focused role for a forward-thinking engineer who thrives at the intersection of AI innovation and real-world delivery. You will own conversation design and live monitoring for voice agent buildout, drive the model ops half of the Bank’s DevOps model, and bring production-grade Python engineering depth alongside AI-specific expertise. Operating across three lifecycle domains — Loans, Deposits, and Payments — you will be a key force in shaping how AI is built and scaled at Customers Bank.

Requirements

5–8 years of software engineering or automation development experience with a strong foundation in Python and demonstrated ability to build and ship production-quality code.
Hands-on experience designing and delivering agentic or automation solutions via custom code, API orchestration, workflow engines, or AI frameworks.
Demonstrated experience with LLM-based development patterns: prompt engineering, tool/function calling, agent frameworks (LangChain, LangGraph, AutoGen), RAG, and NLP-based document or transcript analysis.
Hands-on experience with Microsoft Azure AI workloads — Azure Functions, Logic Apps, Service Bus, Microsoft Foundry, Azure Data Factory, and Azure DevOps.
Strong command of software engineering fundamentals: version control, automated testing, CI/CD, error handling, logging, and documentation.
Experience in financial services or a regulated environment with practical familiarity with compliance, audit trail, model risk, and data governance requirements.

Nice To Haves

Experience building or integrating voice AI / conversational AI solutions.
Experience evaluating AI and automation vendors and contributing to build-vs-buy decisions.
Microsoft Azure certification (Azure Developer Associate, Azure AI Engineer Associate).
Production-grade Python engineering with full SDLC proficiency.
Agent frameworks: LangChain, LangGraph, AutoGen; RAG architecture and NLP pipelines.
Microsoft Azure AI workloads: Functions, Logic Apps, Service Bus, Foundry, Data Factory, DevOps.
OpenAI and Anthropic (Claude) API integration and prompt management.
Voice AI / conversational AI platforms (preferred).
Ability to work with the Microsoft Suite and learn/work with other Customers Bank’s applications.

Responsibilities

Build and scale agentic workflows across Loans, Deposits, and Payments domains.
Design multi-step, autonomous AI pipelines with dynamic decision-making and tool use.
Implement human-in-the-loop (HITL) review gates and confidence scoring thresholds.
Automate structured and semi-structured workflows including exception handling, reconciliations, and regulatory reporting.
Develop, version, and evaluate prompts for OpenAI and Anthropic (Claude) models.
Apply LLM-based development patterns: tool/function calling, agent frameworks (LangChain, LangGraph, AutoGen), and retrieval-augmented generation (RAG).
Build NLP-based intelligent document processing (IDP) pipelines — extraction, classification, and analysis for loan origination and deposit onboarding.
Conduct prompt regression testing and maintain evaluation frameworks to detect output quality degradation.
Monitor live workflow outputs, detect drift or quality degradation, and trigger rollbacks when performance thresholds break.
Manage model and prompt version rollouts in coordination with Cloud Engineering release gates.
Monitor automation performance, troubleshoot production incidents, and continuously optimize deployed solutions.
Design conversation flows, voice persona prompts, and escalation routing rules for the call center voice agent.
Monitor call containment rates, resolution quality, and escalation patterns in production.
Apply NLP-based transcript analysis to continuously improve voice agent performance.
Evaluate AI and automation vendors and frameworks, making informed build-vs-buy recommendations across the partner portfolio.
Partner directly with business SMEs across all three domains to translate operational workflows into AI-executable rules.
Write production-grade Python code with automated testing, error handling, logging, and full documentation.
Design and promote reusable code patterns and architectural standards for enterprise-scale adoption.
Ensure all solutions meet governance, audit trail, model risk, and SOC compliance requirements.
Mentor junior engineers and help establish engineering culture and delivery standards as the team scales.